To say a word correctly, a speaker must retrieve information about the syntactic features (e.g. number of nouns, tense for verbs) and the sound of that particular word from the mental lexicon. We will investigate how the lexicon is organized and represented in the brain. We will approach these issues from the perspective of language production by examining the acquired deficits of word production observed in stroke patients with focal left temporo-parietal lesions and in patients with semantic dementia. Our project will address four questions about lexical processing that are currently under debate in neuroscience: (a) Are there brain structures specific for lexical processing or are the brain structures supporting lexical processing part of a larger circuitry implicated in memory for facts and events (i.e., declaritive memory)? To answer this question we will examine whether word production deficits due to focal left temporo-parietal lesions also show declaritive memory deficits, and whether patients with semantic impairments also have problems in retrieving lexical information. Evidence that these deficits co-occur would suggest that there is a shared circuitry for declaritive memory and lexical processing. (b) How are words processed in the lexicon? Are words like dogs and asked obtained by means of combinatorial processes that assemble the various components of the words (dog+s, ask+ed)? Are there different brain mechanisms for words that do not seem to conform to rules of word formation, such as mice and ate? To address these issues we will examine the selective deficits for producing regular words (dogs, asks) or irregular words (mice, ate) that can be found in brain damaged patients and that have been documented in our preliminary studies. (c) The retrieval of lexical information is likely to involve different stages of processing. What is the nature of theses stages? We will investigate whether different levels of processing are implicated in the retrieval of syntactic and phonological features in word production. (d) What are the neuroanatomical underpinnings of lexical processing? To characterize the brain areas underlying lexical processing, we will carry out an MRI examination of the brain damage responsible for selective word production deficits in stroke patients. In addition to their theoretical relevance for neurocognitive models of lexical processing, our data will be of clinical significance. The results of our studies will provide a more comprehensive picture of word production deficits and memory impairments associated with aphasia and semantic dementia.